The catalytic behaviour of rhodium supported on Al2O3 and TiO2 in synthesis gas reactions

The catalytic behaviour of rhodium supported on Al2O3 and

TiO2 in synthesis gas reactions

Citation for published version (APA):

Blik, van 't, H. F. J., Vis, J. C., Huizinga, T., & Prins, R. (1985). The catalytic behaviour of rhodium supported on

Al2O3 and TiO2 in synthesis gas reactions. Applied Catalysis, 19(2), 405-415.

https://doi.org/10.1016/S0166-9834(00)81761-3

DOI:

10.1016/S0166-9834(00)81761-3

Document status and date:

Published: 01/01/1985

Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be

important differences between the submitted version and the official published version of record. People

interested in the research are advised to contact the author for the final version of the publication, or visit the

DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page

numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne

Take down policy

If you believe that this document breaches copyright please contact us at:

openaccess@tue.nl

providing details and we will investigate your claim.

Elsevier Science Publishers B.V., Amsterdam -Printed in The Netherlands

THE CATALYTIC BEHAVIOUR OF RHODIUM SUPPORTED ON A1203 AND TiO2 IN SYNTHESIS GAS REACTIONS

H.F.J. VAN 'T BLIKa, J.C. VISb, T. HUIZINGAC AND R. PRINSd

Laboratory for Inorganic Chemistry. Eindhoven University of Technology. P-0. Box 513, 5600 MB Eindhoven. The Netherlands. a Present 5600 MB b Present 3130 AC ' Present 1031 CM

a

address: Philips Research Laboratories. P.O. Box 90.000. Eindhoven. The Netherlands.

address: Unilever Research Laboratory, P-0. Box 114. Vlaardingen, The Netherlands.

address: Koninklijke/Shell-Laboratorium, Badhuisweg 3, Amsterdam, The Netherlands.

correspondence should be addressed.

(Received 19 June 1985, accepted '2 August 1985)

ABSTRACT

The catalytic acitivity of Rh/TiO2 and Rh/A1203 catalysts for the hydrogenation of CO at atmospheric pressure and 523 K was investigated. Normal, non-SMSI state Rh/TiO2 as well as Rh/TiO2 catalysts in the SMSI state with varying dispersion were studied. In all cases only hydrocarbons and no oxygenated products were formed. When measured at equal dispersions the activities of Rh/A1203 and non-S&%X Rh/TiO2 catalysts hardly differed. The specific activity of Rh/TiO2 catalysts increased an order of magnitude when changing the dispersion from 1.10 to 0.12. This increase in specific activity was acompanied by an increase in the olefin-to-paraffin ratio, but neither the selectivity to methane nor the probability for chain growth was affected much. Reduction of the Rh/Tio2 catalysts at 773 K decreased their initial activities substantially compared to

reduction at 523 K, whereas the steady-state activities were hardly affected. Apparently the SMSI state is removed to a great extent during CO hydrogenation. The lowering effect of SMSI on activity was a function of dispersion. The effect was much more pronounced for small metal particles, indicating that SMSI might be due to covering. INTRODUCTION

Over the past years rhodium has been gaining importance in catalytic chemistry. Not only is rhodium widely recognized as the best catalyst to promote the reduction of NO in three way catalysts [l-3], it also takes a special place in the conversion of

synthesis gas, since its product range can include oxygenated products (alcohols, aldehydes, acids) besides hydrocarbons [4-111. From literature it is obvious that the character of the support is of main importance for the selectivity and activity of the supported rhodium catalysts in syngas conversion. Nowadays the exotic support titania has enjoyed increased interest. Although titania is a metal oxide support which has been found to strongly inhibit chemisorption of H2 and CO on group VIII metals after reduction at elevated

temperatures, which effect is known as Strong Metal Support Interaction (SMSI) [12], some metals have been found to have their highest activities for CO hydrogenation when dispersed on TiO2. An outstanding example is the titania-supported nickel catalyst [13-161 which shows turnover frequencies one to two orders of magnitude higher than other Ni catalysts. The increase in activity is accompanied by an increase in selectivity to higher-molecular- weight paraffins. Not all group VIII metals show favourable trends in catalytic behaviour when supported on TiO2 [17]. Iron for

instance has a very low activity probably because it becomes easily immersed in the support [la]. With respect to rhodium Solymosi et al. [19] have found that the specific rate of CH4 formation on Rh/Ti02 was more than one order of magnitude higher than that of Rh/MgO and Rh/Si02 catalysts. This effect was attributed to an electronic interaction between the Ti02 and Rh. The authors speculated that Ti02 may influence the bonding and reactivity of species chemisorbed on metal particles.

The anomalous catalytic behaviour of titania-supported catalysts has often been ascribed exclusively to SMSI. However, in many of the studies reported one may doubt if the SMSI effect really was present, because in some cases the reduction temperature used was too low to create SWSI and in other cases a sintering of the metallic particles could not be excluded. In this paper we will present the results of a study of the catalytic behaviour of Rh supported on titania in the normal, non-SMSI-state, as well as in the SMSI-state, and on the more conventional alumina support in the conversion of syngas. A number of rhodium catalysts with different dispersions were used

for this study in order to investigate also the influence of metal particle size on the catalytic behaviour. An extensive characteri- zation study of these catalysts, using temperature programmed reduction and oxidation and transmission electron microscopy, has previously been published [20].

EXPERIMENTAL

TiO2 (anatase. Tioxide Ltd., CLDD 1367, surface area 20 m2 g-l, pore volume 0.5 cm3 g -l) and y-Al 0 (Ketjen.

-1

area 200 m2 g , 23

000-1.5 E, surface pore volume 0.6 cm g-l) were impregnated with aqueous solutions of RhC13. xH20 via the incipient wetness method. In the following the catalysts will be denoted as RT (Rh/Ti02) and RA (Rh/A1203) followed by the metal loading in weight percent. After impregnation the catalysts were dried in air at 355, 375 and

395 K for 2 h successively. followed by direct reduction in flowing H2 at 773 K for 1 h and subsequently passivated and stored for further use.

Catalyst dispersion (percentage exposed) was determined by hydrogen chemicsorption and is expressed as the total amount of hy- drogen atoms adsorbed at room temperature after re-reduction of the RA and RT catalysts at respectively, 773 and 523 K per total amount of rhodium atoms (H/Rh). Note that the titania-supported catalysts were reduced at 773 K during preparation (and thus will have been in the SMSI-state) and passivated and stored afterwards (which removed the SMSI-state [ZO]). Since the RT catalysts were re-reduced at 523 K the H/M values correspond to real dispersions, without inter- ference of sintering effects. For more details we refer to [ZO].

The hydrogenation of CO was investigated in a flow microreactor at 523 K and atmospheric pressure. Two types of mixtures could be passed through the fixed-bed reactor, either a mixture of hydrogen, nitrogen and oxygen, with which the passivated catalysts were pre- treated, or a mixture of hydrogen, carbon monoxide and nitrogen (synthesis gas). Except for oxygen the gases were purified over a BTS column and molecular sieves. The products formed during reaction were analysed on-line by gaschromatography. The GC-equipment

consisted of a Packard Becker 427 gaschromatograph equipped with a 904 TCD detector, and a Pye Unicam gaschromatograph with a flame- ionisation detector. A column packed with Porapack QS, using He as carrier gas, allowed complete separation of C02. ethene, ethane, H20, propene and propane. Using a squalane column with nitrogen as

carrier gas the amount of Cl to CB hydrocarbons could be determined. The amount of catalyst used for an experiment was circa 0.2 g. Prior to the passage of synthesis gas through the reactor, the passivated catalyst was re-reduced in flowing hydrogen for 0.5 h after which the catalyst was cooled under flowing hydrogen to reaction tempera- ture. The alumina-supported catalysts were re-reduced at 773 K and the titania supported catalysts at 523 K or 773 K. The ratio of

H2:CO:N2 in the reaction gas mixture was 2:l:l. The gas hourly space velocity used was 3000 f 100 h-l. In order to measure under

differential conditions the conversion of CO was kept below 10%. If necessary, this was achieved by diluting the catalyst with bare support. The absence of diffusional limitation was confirmed by the method suggested by KSriis and Nowak [21]. Under these reaction

conditions only hydrocarbons were formed and no oxygenates. As the difference in CO concentration between the ingoing and outgoing gas stream could not be accurately determined the amount of CO converted to hydrocarbons was determined in another way. Under reaction conditions the products showed a logarithmic (Flory-Schultz or Anderson [ZZ]) distribution indicating that a constant ratio of propagation and termination rates existed. especially in the molecular range above C2 [23]: In (mole % Cn) = K + n.ln a. The probability of chain growth is a and (l-a) is the probability of chain termination. i.e.. product formation [22.23]. From the experimentally determined C3 to CS hydrocarbons the probability of chain growth a was obtained. Assuming a to be independent of molecular size the concentrations of Cg to Ca, were calculated. The sum of the products (17 nCn) yielded the amount of CO converted to hydrocarbons. The activity is expressed as a turnover frequency: TOFCO = (CO molecules reacted to hydrocarbons)/(surface metal atoms x set). For the number of surface metal atoms the dispersion H/Rh was used.

RESULTS AND DISCUSSION

The activities of all catalysts in the H2 + CO reaction decreased with time on stream. Two deactivation regions could be distinguished, in the beginning of the reaction a fast decline of the reaction rate took place, while after circa 4 h the activity decreased in a more moderate way. Therefore. we define the "initial activity" as the activity obtained by extrapolating the function from the first region (t < 4 h) to reaction time zero, and the "steady-state

activity" as the activity obtained by extrapolating the second region (t > 4 h) to time zero. The activity and selectivity data obtained are summarized in Table 1. Besides the initial and steady-state turnover frequencies (expressed in molecules of CO converted per rhodium surface atom per set), also the dispersions (H/Rh) and the methane selectivity (ScH4 1. the probability of chain growth (a)

TABLE 1

Selectivities and activities obtained at P = 100 kPa, H2:CO:N2 = 2:l:l. and T = 523 K.

INITIAL STEADY-STATE Catalyst H/Rhf T°FCO T°FCO Sg CH4 oh q/c 3 x 10 2 x 10 2 RA 2.3a 1.53 0.31 0.17 0.59 0.44 0.80 RA 4.6 0.96 0.63 0.28 0.61 0.45 1.18 RA 8.5 0.81 0.82 0.35 0.65 0.45 1.56 RA 11.6 0.67 0.73 0.31 0.63 0.45 2.52 RT 0.3b 1.10 0.45 0.10 0.50 0.45 1.50 RT 0.7 0.61 0.74 0.34 0.55 0.51 2.07 RT l.oc 0.40 2.10 0.47 0.49 0.48 2.52 RT 2.0 0.35 2.67 0.58 0.48 0.43 3.41 RT 3.2 0.22 4.11 1.48 0.48 0.42 2.11 RT 8.1d 0.12 4.06 1.43 0.50 0.42 6.50 RT 0.4e 1.10 0.01 0.01 0.48 0.40 1.20 RT 0.7 0.61 0.08 0.07 0.47 0.43 1.95 RT 1.0 0.40 0.40 0.20 0.43 0.42 2.31 RT 2.0 0.35 0.78 0.33 0.48 0.41 3.06 RT 3.2 0.22 1.38 0.91 0.53 0.40 2.05 RT 8.1 0.12 1.11 0.83 0.44 0.42 5.60 a: b: C: d: e: f: 4: h:

The catalysts were re-reduced at 773 K. The catalysts were re-reduced at 523 K. =co = 103 kJ mol-' for RT 1.0.

EC0 = 81 kJ mol-' for RT 8.1.

The catalysts were re-reduced at 773 K.

Based on Hz adsorption at room temperature after re-reduction of the M/A1203 and N/Ti02 catalysts at respectively, 773 and 473 K. 'CH4 = selectivity to methane.

RA and RT in non-SMSI-state

First we will focus our attention to the RA series and RT series

in the non-SMSI-state. Table 1 clearly demonstrates that there

exists a correlation between specific activity and dispersion. The

rate of CO hydrogenation increased by about a factor of 2.5 as the

dispersion of Rh on alumina decreased from 1.53 to 0.67. For the RT

series the activity increased one order of magnitude going from a

dispersion of 1.10 to 0.12. The increase in activity was accompanied

by an increase in the olefin-to-paraffin ratio. Neither the select-

ivity to methane nor the probability of chain growth was greatly

affected. The TOFCO Is in steady-state for the RA series and the RT

series are presented in Figure 1 as a function of dispersion. The

resemblance between the TOFCO Is of the alumina- and titania-

supported rhodium catalysts is clear. Both systems have the same

variation of activity with dispersion and the catalysts with the

same H/Rh value have, within the accuracy of the measurements, the

same turnover frequency. This leads to the conclusion that one must

pay attention to dispersion effects when comparing activities in

H2 + CO reactions of rhodium supported on alumina and titania.

Figure 2 shows that the titania-supported catalysts exhibited a

very rapid decrease in the turnover frequencies for the synthesis of

0 FIT-series re-reduced at 523 K 0 RA-series re- reduced at 773 K

I

0.5 1.0 1.5

H/Rh

Figure 1 Steady-state turnover frequency in the H2 + CO reaction

versus H/Rh (dispersion) for alumina- and titania-supported

x10-2 -Re-reduced at 523K ---- Re-reduced at 773 K

l

Figure 2 Initial and steady-state turnover frequencies in the

Hz + CO reaction versus H/Rh (dispersion) for titania-supported

Rh-catalysts re-reduced at 523 and 773 K. P = 100 kPa, H2:CO:N2

= 2:l:l. T = 523 K.

all products when the dispersion increased above 0.3. An influence

of dispersion on the performance of supported group VIII metals for

the synthesis of hydrocarbons via CO hydrogenation has been reported

previously [24-271. In a detailed study Kellner and Bell 1271 reported

a dramatic decrease in the specific activity for Ru/A1203 catalysts

with dispersions above 0.7. The decrease in activity was accompanied

by a slight decrease in the probability of chain growth and a rapid

decrease in the olefin-to-paraffin ratio. Vannice [25] and

Bartholomew et al. [26] observed that the methanation activity of

Ni catalysts decreased with increasing dispersion. By contrast the

specific activity for methanation of Pt catalysts was found to

increase slightly with increasing dispersion, while no clear cut

dependence on Pd particle size was observed [24].

Several explanations for the decrease in specific activity with

increasing dispersion can be given. One explanation. originally

particles may be changed. Theoretical studies [29] of the electronic properties of small metal particles show that deviations from the properties of bulk metal occur, primarily, for crystallites

smaller than about 20 Y? . This critical size corresponds to a dis- persion of about 0.5. The observed H/Rh value of 0.3. at which a decline of the specific activity takes place, is of the right order of magnitude. A small difference in the electronic properties

between large and small rhodium particles follows from XPS measurements [30]. The observed Rh 3d5,2 electron binding

energies for RT 8.1 (the least dispersed catalyst) and for RT 0.3 (the most dispersed catalyst) after in situ reduction at 483 K are 307.05 and 307.25 eV. respectively. (This small variation in binding energy is most likely caused by differences in the extra-atomic relaxation of metal particles of different sizes. In small particles there is less effective screening of the core holes created during photoemission). This indicates a small change in the electronic properties going from a rhodium loading of 8.1 to 0.3 per cent by weight.

Another explanation for the decrease in specific activity with increasing dispersion might be that the fraction of sites which are most active for the hydrogenation of CO decreases with increasing dispersion. Such a trend would be expected if Rh atoms at the faces of crystallites were more active than those at the edges or corners [31,32]. The apparent activation energy EC0 for RT 8.1 is 81 kJ mol-' which is significantly lower than the activation energy for RT 1.0

-1

which is 103 kJ mol . This result is in accordance with the model because it indicates a change in nature of the active sites with dispersion. This indicates that the nature of the active sites changes with dispersion, but can still be explained by both models.

A completely different explanation is that small particles are much more reactive and therefore prone to deactivation. Larger metal particles have surfaces which deactivate less quickly and on which hydrogen can Still adsorb. The resulting hydrogen atoms may diffuse to more reactive surfaces or sites and raise their steady state activities. The different hydrogenation capabilities of different metals might then explain the difference in activity-dispersion behaviour between Rh. RU and Ni on one hand. and Pt on the other hand.

Comparison of RT series in non-SMSI-state and SMSI-state

The activity and selectivity data obtained by measurements on the Rh/TiO2 catalysts after re-reduction at 773 K are also given

Figure 3 Ratio of initial turnover frequencies in the H2 + CO

reaction of Rh/TiO2 catalysts re-reduced at 773 and 523 K as a

function of H/Rh (dispersion). P = 100 kPa, H2:CO:N2 = 2:l:l.

T = 523 K.

in Table 1. Note that the inluences of dispersion on the catalytic

properties of the RT-series in the SMSI-state and in the non-SMSI-

state are comparable. Significant differences between the catalysts

after a low and high temperature re-reduction appear to be present

in the specific activities only. Neither the selectivity for methane

nor the hydrogenation capacity is affected, whereas the probability

for chain growth tends to decrease somewhat.

Plots of initial and steady-state TOFCOVs as a function of

dispersion for the RT series after low and high temperature re-

reduction are presented in Figure 2. It is obvious that the initial

specific activity is affected most by the SMSI. The effect on the

steady-state activity is, however. moderate. As suggested in the

previous section, this result indicates that SMSI is indeed

destroyed during reaction. if not completely then at least to a

great extent. We consider water. formed during reaction. to be

responsible for this. The difference in the initial activity

between a catalyst re-reduced at low temperature (LT) and the same

catalyst re-reduced at high temperature (HT) and TOFCO initial (LT)

is a measure for SMSI with a ratio of zero indicating a complete

complete lack of SMSI. In order to investigate the influence of dispersion on SMSI we plotted TOFCO initial (HT)/TOFCO initial (LT) versus H/Rh in Figure 3. This figure clearly indicates that SMSI is more pronounced for the dispersed systems. An effect of crystallite size on the onset of SMSI has previously been reported by Ko et al. [33]_ They found that with titania-supported nickel catalysts more severe reduction conditions were necessary to induce strong

metal-support interactions when the nickel crystallites were larger. Two explanations for SMSI have acquired most support over the years. Originally charge transfer from the support to the metal, with a concurrent change in the electron properties of the metal was favoured [12.34]. Lately covering of the metal by suboxides of TiO2 is thought to explain the reduced catalytic activity [35-371. We observed that in situ reduction of Rh/TiO2 at 523 K OK 823 K led to essentially the same Rh 3d5/2 binding energy [30]. This indicates that SMSI is probably not caused by a charge transfer from the support to the metal or vice versa. A plausible explanation for the decrease of the ratio TOFCO initial (HT)/TOFCo initial (LT) with dispersion might be that SMSI is caused by covering. It is obvious that small particles can more easily be encapsulated than larger ones and that this leads to a more pronounced suppression of the initial activity for highly dispersed systems.

ACKNOWLEDGEMENT

This study was supported by the Netherlands Foundation for Chemical Research (SON) with financial aid from the Netherlands Organisation for the Advancement of Pure Research (ZWO).

REFERENCES

1 K.C. Taylor, Catalysis, Science and Technology, vol. 5, J.R. Anderson and M. BoUdart. Eds., Springer Verlag. Berlin (1984).

119.

2 T.P. Kobylinski and B.W. Taylor, J. Catal.. 33 (1974) 376. 3 H.C. Yao. Y.-F. Yao and K. Otto. J. Catal.. 56 (1979) 21. 4 M. Ichikawa. Bull. Chem. Sot. Jap.. 51 (1978) 2268. 5 M. Ichikawa. Bull. Chem. Sot. Jap., 51 (1978) 2273. 6 E.I. Leupold, H.-J. Schmidt, F. Wunder. H.-J. Arpe and

H. Hachenberg. E.P. 0 010 295 Al.

7 F.A. Wunder, H.-J. Arpe, E.I. Leupold and H.-J. Schmidt, Ger. Offen. 28 14 427.

8 W.J. Bartley and T-P. Wilson, Eur. Pat. Appl. 0 021 443. 9 D.G. Castner. R.L. Blackadar and G.A. Somorjai, J. Catal., 66

(1980) 257.

10 P.R. Watson and G.A. Somorjai, J. Catal., 72 (1981) 347. 11 P.R. Watson and G.A. Somorjai, J. Catal., 74 (1982) 282.

12 S.J. Tauster. S.C. Fung and R.L. Garten. J. Am. Chem. Sot.. 100 (1978) 170.

14 M.A. Vannice and R.L. Garten, J. Catal., 66 (1980) 242. 15 M.A. Vannice and J. Vasco-Jara, Stud. Surf. Sci and Catal.,

vol. 11, B. Imelik et al.,Eds., Elsevier, Amsterdam (1982). 185. 16 R. Burch and A.R. Plambard. Stud. Surf. Sci and Catal.. vol. 11,

B. Imelik et al., Eds.. Elsevier, Amsterdam (1982). 193. 17 M.A. Vannice, J. Catal., 74 (1982) 199.

18 B.J. Tatarchuk and J.A. Dumesic, J. Catal., 70 (1981) 308. 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

F. Solymosi. I. Tombacz and M. Kocsis, J. Catal., 75 (1982) 78. J.C. Vis, H.F.J. van It Blik, T. Huizinga, J. van Grondelle and R. Prins, J. Catal., 95 (1985).

R.M. Kords and E.J. Nowak. Chem. Eng. Sci.. 22 (1967) 470. R.B. Anderson, Catalysis, vol. 4. P.H. Emmett. Ed., Reinhold, New York (1956). 257.

G. Henrici-Olive and S. Olive, Angew. Chem. Int. Ed. Engl.. 15 (1976) 136.

M-A. Vannice, J. Catal., 40 (1975) 129. M.A. Vannice. J. Catal., 44 (1976) 152.

C.H. Bartholomew, R.B. Pannell and J.L. Butler, J. Catal., 65 (1980) 335.

C.S. Kellner and A.T. Bell, J. Catal.. 75 (1982) 251. D.C. King, J. Catal., 51 (1978) 386.

J.R. Anderson, Structure of Metallic Catalysts, Academic Press, New York (1975).

T. Huixinga. H.P.J. van It Blik, J.C. Vis and R. Prins. Surf. Sci., 135 (1983) 580.

R.T.K. Baker, E.B. Prestridge and R.L. Garten. J. Catal., 59 (1979) 293.

D.J.C. Yates, L.L. Murrell and E.B. Prestridge, J. Catal.. 57 (1979) 41.

E.I. Ko, S. Winston and C. Woo. J. Chem. Sot., Chem. Commun., (1982) 740.

J.A. Horsley. J. Am. Chem. Sot., 101 (1979) 2870.

P. Meriaudeau, J.P. Dutel, M. Dufaux and C. Naccache. Stud. Surf. Sci. Catal.. 11 (1982) 95.

J. Santos. J. Phillips and J.A. Dumesic, J. Catal.. 81 (1983) 147.